Gene Transfer Establishes Primacy of Striated Vs. Smooth Muscle Sarcoglycan Complex in Limb-Girdle Muscular Dystrophy

Home , Sarcospan, SGCB, Tibialis anterior muscle

Gene transfer establishes primacy of striated vs. smooth muscle sarcoglycan complex in limb-girdle muscular dystrophy

Madeleine Durbeej*, Shanna M. Sawatzki*, Rita Barresi*, Kathleen M. Schmainda†, Vale´ rie Allamand*, Daniel E. Michele*, and Kevin P. Campbell*‡

*Howard Hughes Medical Institute, Department of Physiology and Biophysics, Department of Neurology, University of Iowa, Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA 52242-1101; and †Department of Radiology and Biophysics Research Institute, Medical College of Wisconsin, Milwaukee, WI 53226

Edited by Gerald D. Fischbach, Columbia University College of Physicians and Surgeons, New York, NY, and approved May 19, 2003 (received for review December 16, 2002) Limb-girdle muscular dystrophy types 2E and F are characterized by muscle, ␣-SG deficiency does not lead to cardiomyopathy. Also, skeletal muscle weakness and often cardiomyopathy and are due the dystrophic phenotype appears less severe in ␣-SG-deficient to mutations in the genes encoding ␤- and ␦-sarcoglycan. We mice compared to the muscular dystrophy that develops in ␤- and previously demonstrated that loss of sarcoglycans in smooth mus- ␦-SG-deficient mice. Mice lacking ␤- and ␦-SG display large cle leads to constrictions of the microvasculature that contributes areas of necrosis as the predominant characteristic feature to the cardiac phenotype. It is unclear how vasculature abnormal- similar to alterations observed in tissue infarcts. In addition, ities affect skeletal muscle. We injected recombinant ␤-or␦- arterial constrictions are detected in skeletal muscle (11, 12). sarcoglycan adenoviruses into skeletal muscles of corresponding Despite uniform gene disruption, the surprising heterogeneity of null mice. We hypothesized that the adenoviruses would not phenotype within muscle groups, such as these focal regions of transduce vascular smooth muscle, and we would only target necrosis surrounded by regions of near normalcy, suggests the skeletal muscle. Indeed, sustained expression of intact sarcoglycan– possibility that vascular abnormalities are playing a primary role sarcospan complex was noted at the sarcolemma, neuromuscular in these muscular dystrophies. However, the hypothesis that junction, myotendinous junction, and in peripheral nerve, but not vascular dysfunction directly initiates muscular dystrophy has in vascular smooth muscle. Gene transfer of the corresponding remained untested (11, 12). To address this question, we injected deleted sarcoglycan gene preserved sarcolemmal integrity, pre- recombinant ␤-or␦-SG adenoviruses into corresponding null vented pathological dystrophy and hypertrophy, and protected mice (Sgcb- or Sgcd-null, respectively). We hypothesized that the against exercised-induced damage. We conclude that vascular adenoviruses would not cross the perimysium, in which blood ␤ ␦ dysfunction is not a primary cause of - and -sarcoglycan-deﬁ- vessels run, and thus we would only target skeletal muscle. The cient muscular dystrophy. In addition, we show successful func- sarcoglycans were successfully restored at the sarcolemma, neu- tional rescue of entire muscles after adenovirus-mediated gene romuscular junction (NMJ), myotendinous junction (MTJ), and delivery. Thus, virus-mediated gene transfer of sarcoglycans to peripheral nerve, but not in smooth muscle of vessels within the skeletal muscle in combination with pharmacological prevention of skeletal muscle. Despite lack of restoration in vascular smooth cardiomyopathy constitute promising therapeutic strategies for muscle, the dystrophic phenotype was prevented, as demon- limb-girdle muscular dystrophies. strated by improved skeletal muscle morphology, restored sarcolemmal integrity, and reduction of muscle mass. Also, injec- he sarcoglycans (␣, ␤, ␥, and ␦-SG) are localized at the tion of ␦-SG adenovirus was able to prevent the development of Tsarcolemma of muscle fibers and form, along with sarcospan, exercise-induced injury. We conclude that (i) disruption of the a subcomplex within the larger dystrophin–glycoprotein com- DGC in vascular smooth muscle is not a primary cause of plex (DGC) (1–3). The DGC is further composed of the muscular dystrophy; (ii) primary mutations in ␤- and ␦-SG can intracellularly located protein dystrophin, which binds to the be functionally corrected in vivo; and (iii) structures such as ␤ transmembrane protein -dystroglycan, which is associated with NMJ, MTJ, and peripheral nerve that are affected in muscular ␣ ␣ the peripheral protein -dystroglycan. -Dystroglycan, in turn, dystrophy and other neuromuscular diseases can be targets for binds to laminin-2, present in the skeletal muscle basement adenoviral-mediated gene delivery. membrane (4, 5). Thus, the DGC provides a structural link between the extracellular matrix and the intracellular cytoskel- Methods eton of muscle cells. The importance of the sarcoglycans for Animals. We previously reported the generation of Sgcb-null normal skeletal muscle function is underscored by the findings ␥ ␣ ␤ ␦ mice, lines Q94 and Q283 (12). Here, we have also analyzed that mutations in the genes for -, -, -, and -SG cause Sgcb-null mice from a third clone, Q201. Sgcd-null mice, line I74 limb-girdle muscular dystrophy types 2C-F, respectively. Fur- were also described (11). All mice were on hybrid C57BL͞6– thermore, one mutant sarcoglycan isoform causes the loss or the 129-SvJ background. Colonies were maintained in the Animal reduction at the muscle cell membrane of all of the other Care Unit of the University of Iowa College of Medicine, sarcoglycan proteins in the DGC (6). according to the animal care guidelines. The Animal Care and In striated muscle, the sarcoglycan complex is composed of ␣-, Use Review Committees of the University of Iowa and the ␤-, ␥-, and ␦-SG. In addition, a second sarcoglycan complex in Medical College of Wisconsin approved all animal procedures. which ␧-SG replaces ␣-SG is also expressed in striated muscle and is predominant in smooth muscle (7–10). Loss of the sarcoglycans in vascular smooth muscle results in perturbed This paper was submitted directly (Track II) to the PNAS ofﬁce. vascular function with the presence of vascular constrictions Abbreviations: SG, sarcoglycan; DGC, dystrophin–glycoprotein complex; NMJ, neuromus- leading to intermittent ischemic-like events, which initiate the cular junction; MTJ, myotendinous junction; EBD, Evans blue dye; nNOS, neuronal nitric development of cardiomyopathy in mouse models of ␤- and ␦-SG oxide synthase. deficiency (11–13). Because ␣-SG is not expressed in smooth ‡To whom correspondence should be addressed. E-mail: [email protected].

8910–8915 ͉ PNAS ͉ July 22, 2003 ͉ vol. 100 ͉ no. 15 www.pnas.org͞cgi͞doi͞10.1073͞pnas.1537554100 Downloaded by guest on September 26, 2021 Recombinant Adenoviral Vectors. Human ␤- and ␦-SG sequences were amplified by PCR and subcloned into the pAdCMVpA adenovirus shuttle vector. The ␤- and ␦-SG constructs were then incorporated into an adenovirus vector through standard methods of homologous recombination with Ad5 backbone dl309 by the University of Iowa Gene Transfer Vector Core. First- generation recombinant viruses were purified as described (12, 14).

Adenoviral Vector Administration. Sgcb- and Sgcd-null pups (2- to 4-day-old) were anesthetized via hypothermia by placement on ice-cooled aluminum foil for 1–2 min. Infectious particles (1–2 ϫ 109) of Ad5 cytomegalovirus containing human ␤-or␦-SG cDNA diluted in a final volume of 10 ␮l in 0.9% NaCl were injected percutaneously into each skeletal muscle of the corresponding null mice by using an insulin syringe. Pups were reintroduced to the mother and kept in quarantine for 5 days. All pups survived after injections.

Immunohistochemical Analysis. Stainings were performed as described (12) with Abs against ␤-dystroglycan (15); ␣-, ␤-, ␥-, and ␦-SG (8, 14, 16, 17); sarcospan (18), and neuronal nitric oxide synthase (nNOS) (19). The smooth muscle actin Ab was from Sigma. Secondary Abs were conjugated with FITC or Cy3 (Jackson ImmunoResearch) or Alexa Fluor 488 (Molecular Probes). For staining of NMJs, samples were simultaneously incubated with FITC-conjugated ␣-bungarotoxin (Molecular Probes). Sections were observed under an MRC-600 laser scanning confocal microscope (Bio-Rad) or a Leica DMRXA MEDICAL SCIENCES fluorescence microscope (Leica, Deerfield, IL).

Biochemical Analysis. Glycoprotein preparations were performed as described (12) and analyzed by using Abs against ␣-SG (␣Sarc͞20A6); ␤-SG (␤Sarc͞5B1); ␣-dystroglycan (IIH6) (20); ␣2 subunit of the dihydropyridine receptor (21); ␦-SG (14) and sarcospan (18).

MRI. Before imaging, mice were anesthetized with 87.5 mg͞kg ketamine and 12.5 mg͞kg xylazine and surgically instrumented Fig. 1. Restoration of the sarcoglycan–sarcospan complex and nNOS at the with vein catheters for the administration of gadolinium (Ga- sarcolemma upon adenovirus-mediated expression of ␤- and ␦-SG. (a) Quad- dodiamide, Omniscan; Nycomed, Oslo). All experiments were riceps muscle cryosections from WT, Sgcb-null, Sgcb-null mice injected with performed at the Medical College of Wisconsin on a 3-T Bruker recombinant ␤-SG adenovirus, Sgcd-null, and Sgcd-null mice injected with Medspec MR imaging system (Bruker, Billerica, MA). Mice recombinant ␦-SG adenovirus were stained with Abs against ␣-SG, ␤-SG, ␥-SG, ␦ were placed in a 10-cm three-axis local gradient coil, together -SG, sarcospan (SSPN), and nNOS. (b) Glycoprotein preparations from quad- ϫ ͞ riceps muscle of WT, Sgcd-null, and Sgcd-null mice injected with ␦-SG adeno- witha2 4-cm quadrature transmit receive radiofrequency virus were analyzed after 16 days by SDS͞PAGE and immunoblotting by using surface coil. After obtaining sagittal scout images, five to seven Abs against ␣-SG, ␤-SG, ␥-SG, ␦-SG, and sarcospan (SSPN) as indicated. Abs 2-mm axial slices were chosen to image the tibialis anterior. against ␣2-dihydropyridine receptor (DHPR) demonstrate equal loading. T1-weighted spin-echo images were acquired after 20 min after administration of 0.2 mmol͞kg gadolinium. anterior, quadriceps femoris, hamstring (gracilis, semitendino- Treadmill Exercise. Animals were exercised by using the Om- sus, semimembranosus, biceps femoris), and calf (gastrocnemius nipacer treadmill (model LC4͞M-MGA͞AT, Accuscan Instru- and soleus) muscles with adenoviruses containing the human ␤- ments, Columbus, OH). Eight-week-old WT, Sgcd-null, and or ␦-SG cDNAs, respectively, under the control of the cytomeg- Sgcd-null mice injected with ␦-SG adenovirus were exercised at alovirus promoter. Renewed expression of the corresponding a 15° downward angle with increasing speed up to 19 m͞min, sarcoglycan in tibialis anterior, quadriceps femoris, hamstring, allowed to rest, and then the procedure was repeated. All mice and calf muscles was coincident with the rescue of the entire were injected with Evans blue dye (EBD) i.p. 5 h before exercise. sarcoglycan–sarcospan complex to the sarcolemma based on ␣-, Mice were killed 24 h after exercise, and sections of muscles were ␤-, ␥-, ␦-SG, and sarcospan-specific immunofluorescence (Fig. studied for EBD uptake and ␤-dystroglycan and ␣-SG expres- 1a and data not shown). This result is in accordance with sion. Quantification of EBD-positive stained areas in sections of previous data from gene transfer experiments of ␤-SG in skeletal muscle was done by using the SCION image program Sgcb-null mice (12). The DGC also directly or indirectly binds (Scion, Frederick, MD). The percentage of positive-stained accessory proteins, including nNOS and Grb2 (22, 23). Sarco- areas was calculated by dividing the area of staining by the total lemma localization of nNOS was severely reduced in Sgcb- and area of the analyzed skeletal muscle section. Sgcd-null muscle but maintained at the sarcolemma after gene transfer (Fig. 1a). Results In agreement with the immunofluorescence data, Sgcd-null Restoration of DGC upon ␤- and ␦-SG Gene Transfer. We injected 2- muscle injected with ␦-SG adenovirus exhibited renewed ex- to 4-day-old Sgcb- or Sgcd-null pups directly into the tibialis pression of the sarcoglycan–sarcospan complex as revealed

Durbeej et al. PNAS ͉ July 22, 2003 ͉ vol. 100 ͉ no. 15 ͉ 8911 Downloaded by guest on September 26, 2021 by immunoblot analysis of glycoprotein preparations of quadriceps muscle (Fig. 1b). The association between the sarcoglycan complex and ␣-dystroglycan was also reestablished (data not shown). Together, these results provide strong evidence that the integrity of the DGC has been restored in Sgcb- and Sgcd-null muscle expressing the corresponding ␤- and ␦-SG adenoviruses.

Reexpression of Sarcoglycans into Neuromuscular Junctions, Myoten- dinous Junctions, and Schwann Cells. We have previously shown that the sarcoglycans can be restored at the sarcolemma after adenovirus mediated ␣-SG gene transfer to ␣-SG-deficient mice (24). However, it has remained untested whether gene transfer can target specialized muscle membranes and other cell-types within skeletal muscle. For example, the sarcoglycans are normally expressed at the NMJ where they constitute a subcomplex within the utrophin–glycoprotein complex (25). Double stainings with FITC-coupled ␣-bungarotoxin, which binds acetylcho- line receptors, revealed that the sarcoglycans were restored to the NMJs after ␤-or␦-SG gene transfer (Fig. 2a). Likewise, the sarcoglycans were re-established at the MTJ (Fig. 2b). A sarcoglycan complex is also present in the peripheral nerve and this complex is absent in Sgcb- and Sgcd-null mice (ref. 26; Fig. 2c). After ␤-or␦-SG gene transfer, the sarcoglycans were restored at the Schwann cell plasma membrane of the corresponding null mice (Fig. 2c).

The Vascular Smooth Muscle Sarcoglycan Complex Is Not Reconsti- tuted on Gene Transfer. The smooth muscle sarcoglycan complex (␧-, ␤-, ␥-, and ␦-SG) is normally expressed in vascular smooth muscle where it colocalizes with smooth muscle actin (Fig. 2d). However, it has remained unclear whether smooth muscle in blood vessels can be transduced by i.m. injected adenoviral vectors. On ␤-or␦-SG gene transfer, the sarcoglycans were only reconstituted at the sarcolemma but not in the vascular smooth muscle, as revealed by double staining with smooth muscle actin (Fig. 2d).

Sarcolemmal Integrity Is Restored on ␦-SG Gene Transfer. After gene transfer (10 wk), sarcolemmal membrane permeability was assessed in vivo by using MRI of infused gadolinium, a contrast agent that binds reversibly to intracellular serum albumin. Absence of the sarcoglycan complex results in altered membrane permeability, which allows extracellular serum albumin and Fig. 2. Sarcoglycan expression is restored at the NMJ, MTJ, and Schwann other serum proteins to cross into skeletal muscle fibers (11, 12, cells, but not vasculature, on ␤- and ␦-SG gene transfer. (a) Sections from 17). T1-weighted spin-echo images obtained from the tibialis skeletal muscle containing NMJ of WT, Sgcb-null, Sgcb-null mice injected with anterior muscles of noninjected Sgcd-null mice showed extensive recombinant ␤-SG adenovirus, Sgcd-null, and Sgcd-null mice injected with areas of muscle damage as detected by uptake of the contrast recombinant ␦-SG adenovirus were doubly stained with ␣-SG Ab (␣-SG, red) agent. Interestingly, there was significantly less uptake in the and ﬂuorescein ␣-bungarotoxin (␣-BTX, green). (b) Sections from skeletal injected muscle in all injected animals (Fig. 3). Together, these muscle of WT, Sgcb-null, Sgcb-null mice injected with recombinant ␤-SG data demonstrate that the compromised sarcolemmal integrity adenovirus, Sgcd-null, and Sgcd-null mice injected with recombinant ␦-SG in Sgcd-null mice can be reversed by restoration of the sarco- adenovirus were stained with ␤-SG Abs. Arrows indicate MTJs. (c) Sections of glycan complex. Similar results were previously obtained by skeletal muscle containing peripheral nerve of WT, Sgcb-null, Sgcb-null mice ␣ ␣ injected with recombinant ␤-SG adenovirus, Sgcd-null, and Sgcd-null mice using adenovirus-mediated -SG gene transfer to -SG- injected with recombinant ␦-SG adenovirus were stained with Abs against deficient mice (24). ␤-SG. (d) Sections from skeletal muscle containing smooth muscle of WT, Sgcb-null, and Sgcd-null mice injected with the corresponding sarcoglycan Human ␤- and ␦-SG Gene Transfer Prevents Dystrophic Pathology of adenovirus were doubly stained with ␤-SG (␤-SG, red) or ␦-SG (␦-SG, red) and Skeletal Muscles. We next examined the morphology of injected smooth muscle actin (actin, green) Abs. and contralateral noninjected muscles of 35 Sgcb- or Sgcd-null mice 4–35 wk after injection. Histological features of the dystrophic muscle in Sgcb- and Sgcd-null mice include internally duction efficiency of calf muscles was also Ϸ40–80% (Fig. 5 and placed nuclei, large areas of necrosis, and dystrophic calcification data not shown). The dystrophic phenotype was prevented in (11, 12). One single injection of adenovirus into the tibialis transduced muscles, exhibited by absence of necrotic areas and anterior and hamstring muscles typically resulted in a 70–100% dystrophic calcifications (Fig. 4). In addition, significantly fewer transduction of muscle fibers as revealed by immunofluores- central nuclei were noted in transduced muscles (Figs. 4b and cence (Figs. 4a and 5 and data not shown). Injections into 6a). For several muscles of Sgcb- and Sgcd-null mice (Ϸ30-wk- quadriceps femoris muscle resulted in transduction of vastus old), compared to WT mice, the masses were 30–95% greater lateralis, but not rectus femoris (12, 24) with an efficiency of (Fig. 6b and data not shown), most likely due to pathological transduction of Ϸ40–80% (Fig. 5 and data not shown). Trans- hypertrophy. Because high transduction efficacy (70–100%) was

8912 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.1537554100 Durbeej et al. Downloaded by guest on September 26, 2021 Fig. 3. Adenovirus-mediated gene transfer of ␦-SG prevents uptake of Fig. 4. Adenovirus-mediated expression of ␤- and ␦-SG protects the tibialis gadolinium in tibialis anterior muscle. Contrast agent-enhanced MRI of WT anterior muscle from dystrophic damage. (a Left) Severe dystrophic changes

(Top), noninjected (Middle), and right-leg-injected Sgcd-null (Bottom) mice. including central nucleation, necrosis (horizontal arrows), and dystrophic MEDICAL SCIENCES Cross section images through the left (L) and right (R) tibialis anterior muscles calcification (vertical arrows) were detected by hematoxylin and eosin stain- were taken 20 min after injection of gadolinium. T denotes the tail of the ing of cryosections of noninjected tibialis anterior muscles of Sgcb-null and mouse. Note lack of uptake of contrast agent in WT muscles; dramatic uptake Sgcd-null mice. (Center) Contralateral tibialis anterior muscles 1 mo after i.m. in Sgcd-null muscles (white areas) and major decrease in enhancement levels injections of ␤-or␦-SG adenoviruses. Note that transduced fibers were pro- in right-side-injected muscles. tected from dystrophic damage. (Right) Serial sections of injected muscles stained with a polyclonal Ab against ␣-SG. (b) Higher magnification of muscles in a. exhibited in tibialis anterior and hamstring muscles, we analyzed the mass of these muscles after gene transfer. Indeed, the mass of transduced Sgcb- and Sgcd-null tibialis anterior and hamstring and specialized muscle membranes (NMJ and MTJ) within muscles was significantly less than the mass of uninjected Sgcb- skeletal muscle are involved. For example, loss of the sarcogly- and Sgcd-null muscles (Fig. 6b). Importantly, muscle fibers cans in vascular smooth muscle of Sgcb- or Sgcd-null mice is remained transduced for at least 35 wk, and the morphology of hypothesized to be necessary for the development of cardiomy- transduced muscles was significantly preserved (data not shown). opathy characterized by vascular constrictions and increased markers of ischemic damage. Also, vascular constrictions are ␦-SG Gene Transfer Protects Against Exercise-Induced Injury. To present in skeletal muscle of these mice, but it is not clear demonstrate functional benefit conferred by the virus-mediated whether they also contribute to the muscular dystrophy pheno- expression of sarcoglycans, we exercised 8-wk-old WT, Sgcd- and type (11–13). To address this question, we injected recombinant Sgcd-null mice injected with ␦-SG adenovirus in hind limb ␤-or␦-SG adenoviruses into corresponding null mice. We muscles. Skeletal muscles, in particular hamstring, quadriceps, assumed that the inability of adenoviral particles to cross the and calf muscles, of Sgcd-null mice were susceptible to exercise- perimysium (24), in which many arteries run, would prevent induced sarcolemmal injury as evidenced by increased uptake of transduction of vascular smooth muscle. One single injection of EBD (Fig. 5) compared to exercised WT animals (data not ␤-or␦-SG adenovirus into skeletal muscle of neonatal pups shown) and nonexercised Sgcd-null muscle (Fig. 5). Quantitative deficient in ␤-or␦-SG, respectively, resulted in a complete image analysis indicated EBD uptake in hamstring muscles restoration of the skeletal muscle DGC at the sarcolemma and increased from 3.7 to 28.6% upon exercise whereas in quadriceps prevented the dystrophic pathology. The vascular smooth muscle and calf muscles, EBD uptake increased from 1.8 to 18% and DGC, on the other hand, was not restored. Because we were able from 2.0 to 4.1%, respectively. Tibialis anterior muscle of to prevent muscular dystrophy, we conclude that vascular per- Sgcd-null mice was less susceptible to exercise-induced injury turbations do not significantly contribute to the skeletal muscle (EBD uptake increased from 0.2% to 1.8%). Interestingly, ␦-SG pathology. If vascular dysfunction does not play a major role in adenoviral injections were able to dramatically reduce exercise- the pathogenesis of muscular dystrophy of Sgcb- or Sgcd-null induced injury (Fig. 5). The EBD uptake in exercised tibialis mice, why do mice deficient in ␣-SG (with an intact vascular anterior, hamstring, quadriceps, and calf muscles injected with smooth muscle complex) present a less severe muscle phenotype ␦-SG adenovirus was 2.4%, 1.9%, 1.2%, and 0.1%, respectively. compared to mice deficient in ␤-or␦-SG, which have a disrupted smooth muscle sarcoglycan complex? One plausible explanation Discussion is the difference in skeletal muscle expression of ␧-SG between Muscular dystrophy represents a group of skeletal muscle dis- the mutant animals. Mice lacking ␣-SG still express the ␧-SG eases that are characterized by progressive muscle weakness and complex in skeletal muscle, whereas the skeletal muscle ␧-SG wasting. However, not only the muscle fiber per se is affected in complex is grossly perturbed in Sgcb- or Sgcd-null mice (7, 12, muscular dystrophy but also other cell types (smooth muscle) 17 and unpublished data). Thus, the ␧-SG complex may partially

Durbeej et al. PNAS ͉ July 22, 2003 ͉ vol. 100 ͉ no. 15 ͉ 8913 Downloaded by guest on September 26, 2021 Fig. 5. Injection of ␦-SG adenovirus protects against exercised-induced injury. Composite cross-sectional images of tibialis anterior, hamstring, quadriceps, and calf muscles from Sgcd-null mice, exercised Sgcd-null mice, Sgcd-null mice injected with ␦-SG adenovirus (Sgcd-null Ad), and exercised Sgcd-null mice injected with ␦-SG adenovirus. Staining in green denotes ␤-dystroglycan (ﬁrst four columns) or ␣-SG (last column), and staining in red denotes EBD uptake. (Bars ϭ 220 ␮m.)

compensate for the loss of the predominant ␣-SG-containing complex. We herein report successfully reconstituted sarcoglycan com- plexes at the NMJ, MTJ, and peripheral nerves upon gene transfer. The transduction of peripheral nerve is surprising considering the fact that both blood vessels and nerves are located in the perimysium. Nevertheless, it is still possible that the virus can cross the perimysium and failure of transduction of vascular smooth muscle could be due to other factors. An alternative explanation is that the virus cannot cross the perimysium and infects the peripheral nerve by other means, possibly by infection of Schwann cell precursors or retrograde transport from the NMJ. Genes encoding for components of the NMJ are expressed in the subsynaptic nuclei, specialized myonuclei that lie beneath the NMJ (27). It was unclear whether the virus and promoter would target these nuclei. However, the successful restoration of the sarcoglycans at the NMJ suggests successful adenoviral targeting to these specialized myonuclei. Although overt structural defects of the NMJs, MTJs, and peripheral nerves have not yet been reported in sarcoglycan- deficient animals, absence of other DGC components affects these structures (11, 12, 17, 26). Loss of dystroglycan in chimeric mice, for example, results in aberrant synapses in addition to muscular dystrophy (28). Likewise, ␣-dystrobrevin is important for the maturation and maintenance of the neuromuscular synapse and loss of ␣-dystrobrevin also results in muscular dystrophy (29, 30). Absence of integrin ␣7 leads to muscular dystrophy and pathological changes of the myotendinous junctions (31). Finally, lack of laminin ␣2 results in develop- Fig. 6. Adenovirus-mediated expression of ␦- and ␤-SG protects myofibers mental dysmyelination of peripheral nerve in addition to severe from central nucleation and increased muscle mass. (a) Percentage of trans- muscular dystrophy (32–35). Thus, our data on successful duced fibers with centrally located nuclei 2 mo after adenovirus injections into adenoviral-mediated gene delivery to NMJ, MTJ, and periph- Ⅺ f Sgcd-null tibialis anterior ( ), hamstring ( ), quadriceps ( ), and calf ( ) eral nerve can be adapted to study these DGC components or muscles. *, Difference from Sgcd-null, P Ϸ 0; #, difference from Sgcd-null, P Ͻ 0.0015. (b) Tibialis anterior (Ⅺ) and hamstring (f) muscle masses of 30-wk-old even disparate proteins involved in disorders affecting these membrane specializations and differential cell types. mice. *, Difference from Sgcb-null, P Ͻ 0.026; #, difference from Sgcd-null, P Ͻ 0.0002; †, difference from Sgcb-null and WT, P Ͻ 0.001 and P Ͻ 0.008, To date, there have been several reports on successful ade- respectively; ‡, difference from Sgcd-null, P Ͻ 0.027. Statistical significance novirus-mediated gene transfer resulting in morphologically was examined by using Student’s t test. rescued muscles in various animal models of muscular dystrophy

8914 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.1537554100 Durbeej et al. Downloaded by guest on September 26, 2021 (14, 17, 24, 36). However, few reports have demonstrated any mice are greater. The decreased muscle masses of injected Sgcb- functional benefits of muscle gene transfer. A major functional null and Sgcd-null mice thus suggest that injected mice produce defect in dystrophic muscle of mdx mice is muscle damage and normal force levels. Together, we provide the first evidence for sarcolemma disruptions caused by eccentric exercise (37). Also, successful functional rescue of entire muscles after adenoviral- treadmill exercise initiates the development of cardiac muscle mediated gene-delivery. necrosis in young Sgcd-null mice (11). Yet, the effects of exercise In conclusion, we have established that vascular perturbation on ␤-or␦-SG-deficient skeletal muscle had been unclear. In in skeletal muscle is not a primary cause of ␤- and ␦-SG-deficient Sgcd-null mice, our downhill running protocol produced a large muscular dystrophy. In cardiac muscle, however, the presence of increase in percentage of dye permeable muscle fibers, com- the vascular constrictions and ischemic-like injury appears to pared to nonexercised Sgcd-null mice and exercised normal initiate the development of cardiomyopathy. We have demon- ␤ ␦ mice. Thus, muscle fibers of Sgcd-null mice are more vulnerable strated that cardiomyopathy in - and -SG-deficient mice can to exercise induced sarcolemma damage than those of normal be prevented by using verapamil, a calcium channel blocker (13). mice. Importantly, we report effective prevention of exercised Thus, early virus-mediated gene transfer of sarcoglycans to induced injury and sarcolemma damage by adenovirus-mediated skeletal muscle in combination with pharmacological interven- gene transfer. The muscle masses of 30-wk-old Sgcb-null and tion constitute promising therapeutic strategies for limb-girdle Sgcd-null mice were significantly greater than muscle masses of muscular dystrophies associated with cardiomyopathy. WT mice. Hypertrophy constitutes an adaptation to the contin- We gratefully acknowledge Beverly L. Davidson and Richard D. Ander- uous damage to muscle fibers. Lynch et al. (38) demonstrated son and the Gene Therapy Center Vector Core Facility, supported by that muscle hypertrophy in mdx mice is highly effective in the National Institutes of Health Grant P30DK54759, for their assistance. maintenance of absolute force, although the specific force was This work was supported in part by the Muscular Dystrophy Association. lower than that of the control mice since muscle masses of mdx K.P.C. is an Investigator of the Howard Hughes Medical Institute.

1. Campbell, K. P. & Kahl, S. D. (1989) Nature 338, 259–263. 20. Ervasti, J. M. & Campbell, K. P. (1991) Cell 66, 1121–1131. 2. Ervasti, J. M., Ohlendieck, K., Kahl, S., Gaver, M. G. & Campbell, K. P. (1990) 21. Ohlendieck, K. & Campbell, K. P. (1991) J. Cell Biol. 115, 1685–1694. Nature 345, 315–319. 22. Yang, B., Jung, D., Motto, D., Meyer, J., Koretzky, G. & Campbell, K. P. (1995) 3. Crosbie, R. H., Lebakken, C. S., Holt, K. H., Venzke, D. P., Straub, V., Lee, J. Biol. Chem. 270, 11711–11714. J. C., Grady, R. M., Chamberlain, J. S., Sanes, J. R. & Campbell, K. P. (1999) 23. Brenman, J. E., Chao, D. S., Xia, H., Aldape, K. & Bredt, D. S. (1995) Cell 82, J. Cell Biol. 145, 153–165. 743–752. MEDICAL SCIENCES 4. Ervasti, J. M. & Campbell, K. P. (1993) J. Cell Biol. 122, 809–823. 24. Allamand, V., Donahue, K. M., Straub, V., Davisson, R. L. & Campbell, K. P. 5. Jung, D., Yang, B., Meyer, H. J., Chamberlain, J. S. & Campbell, K. P. (1995) (2000) Gene Ther. 7, 1385–1391. J. Biol. Chem. 270, 273005–27310. 25. Matsumura, K., Ervasti, J. M., Ohlendieck, K., Kahl, S. D. & Campbell, K. P. 6. Cohn, R. D. & Campbell, K. P. (2000) Muscle Nerve 23, 1456–1471. (1992) Nature 360, 588–591. 7. Liu, L. A. & Engvall, E. (1999) J. Biol. Chem. 274, 38171–38176. 26. Imamura, M., Araishi, K., Noguchi, S. & Ozawa, E. (2000) Hum. Mol. Genet. 8. Durbeej, M. & Campbell, K. P. (1999) J. Biol. Chem. 274, 26609–26616. 9. Straub, V., Ettinger, A. J., Durbeej, M., Venzke, D. P., Cutshall, S., Sanes, J. R. 9, 3091–3100. & Campbell, K. P. (1999) J. Biol. Chem. 274, 27989–27996. 27. Schaeffer, L., de Kerchove d’Exaerde, A. & Changeux, J. P. (2001) Neuron 31, 10. Barresi, R., Moore, S. A., Stolle, C. A., Mendell, J. & Campbell, K. P. (2000) 15–22. J. Biol. Chem. 275, 38554–38560. 28. Coˆte´, P. D., Moukhles, H., Lindenbaum, M. & Carbonetto, S. (1999) Nat. 11. Coral-Vazquez, R., Cohn, R. D., Moore, S. A., Hill, J. A., Weiss, R. M., Genet. 23, 338–342. Davisson, R. L., Straub, V., Barresi, R., Bansal, D., Hrstka, R. F., et al. (1999) 29. Grady, R. M., Grange, R. W., Lau, K. S., Maimone, M. M., Nichol, M. C., Stull, Cell 98, 465–474. J. T. & Sanes, J. R. (1999) Nat. Cell Biol. 4, 215–220. 12. Durbeej, M., Cohn, R. D., Hrstka, R. F., Moore, S. A., Allamand, V., Davidson, 30. Grady, R. M., Zhou, H., Cunningham, J. M., Henry, M. D., Campbell, K. P. B. L., Williamson, R. A. & Campbell, K. P. (2000) Mol. Cell 5, 141–151. & Sanes, J. R. (2000) Neuron 25, 279–293. 13. Cohn, R. D., Durbeej, M., Moore, S. A., Coral-Vazquez, R., Prouty, S. & 31. Mayer, U., Saher, G., Fa¨ssler,R., Bornemann, A., Echtermeyer, F., von der Campbell, K. P. (2001) J. Clin. Invest. 107, R1–R7. Mark, H., Miosge, N., Poschl, E. & von der Mark, K. (1997) Nat. Genet. 17, 14. Holt, K. H., Lim, L. E., Straub, V., Venzke, D. P., Duclos, F., Anderson, R. D., 318–323. Davidson, B. A. & Campbell, K. P. (1998) Mol. Cell 1, 841–848. 32. Bradley, W. G. & Jenkinson, M. (1973) J. Neurol. Sci. 18, 227–247. 15. Williamson, R. A., Henry, M. D., Daniels, K. J., Hrstka, R. F., Lee, J. C., 33. Madrid, R. E., Jaros, E., Cullen, M. J. & Bradley, W. G. (1975) Nature 257, Sunada, Y., Ibraghimov-Beskrovnaya, O. & Campbell, K. P. (1997) Hum. Mol. 319–321. Genet. 6, 831–841. 34. Xu, H., Christmas, P., Wu, X. R., Wewer, U. M. & Engvall, E. (1994) Nat. 16. Roberds, S. L., Anderson, R. D., Ibraghimov-Beskrovnaya, O. & Campbell, Genet. 8, 297–302. K. P. (1993) J. Biol. Chem. 268, 23739–23742. 35. Sunada, Y., Bernier, S. M., Utani, A., Yamada, Y. & Campbell, K. P. (1995) 17. Duclos, F., Straub, V., Moore, S. A., Venzke, D. P., Hrstka, R. F., Crosbie, R. H., Durbeej, M., Lebakken, C. S., Ettinger, A. J., van der Meulen, J., et al. Hum. Mol. Genet. 4, 1055–1061. (1998) J. Cell Biol. 142, 1461–1471. 36. Wakefield, P. M., Tinsley, J. M., Wood, M. J., Gilbert, R., Karpati, G. & Davies, 18. Lebakken, C. S., Venzke, D. P., Hrstka, R. F., Consolino, C. M., Faulkner, J. A., K. E. (2000) Gene Ther. 7, 201–204. Williamson, R. A. & Campbell, K. P. (2000) Mol. Cell. Biol. 20, 1669–1677. 37. Brussee, V., Tardif, F. & Tremblay, J. P. (1997) Neuromuscul. Disord. 7, 19. Crosbie, R. H., Straub, V., Yun, H.-Y., Lee, J. C., Rafael, J. A., Chamberlain, 487–492. J. S., Dawson, V. L., Dawson, T. M. & Campbell, K. P. (1998) Hum. Mol. Genet. 38. Lynch, G. S., Hinkle, R. T., Chamberlain, J. S., Brooks, S. V. & Faulkner, J. A. 7, 823–829. (2001) J. Physiol. (London) 535, 591–600.

Durbeej et al. PNAS ͉ July 22, 2003 ͉ vol. 100 ͉ no. 15 ͉ 8915 Downloaded by guest on September 26, 2021